Effects and Control of Pulsation in Gas Measurement

Effects and Control of Pulsation in Gas
Measurement
Introduction
Pulsation created by compressors, flow
control valves, regulators and some piping
configurations are known to cause
significant errors in gas measurement.
In recent years the Pipeline-and Compressor
Research Council (PCRC) now know as
(GMRC) Gas Machinery Research Council, a
subsidiary of the Southern Gas Association,
commissioned and funded various pulsation
research projects at Southwest Research
Institute (SWRI) in San Antonio, Texas.
This research culminated in the publication of
several technical papers, including the April
1987 PCRC report 10.87-3 titled "Pulsation and
Transient-induced Errors at Orifice Meter
Installations" and a report, 'An Assessment of
Technology for Correcting Pulsation Induced
Orifice Flow Measurement" dated November
1991.
The PCRC sponsored research programs
concluded that pulsation induced measurement
errors fall into two broad categories:
1) Primary Element Error: which includes
square root averaging error (SRE),
inertial errors, and shifts in the orifice
coefficient?
2) Secondary Element Errors: which
consist of gauge line distortion and
gauge line shift, together commonly
referred to as gauge line error (GLE).
This paper will concentrate on methods to
determine Square Root Error, Gauge Line
Error, and suggest several techniques to reduce
these effects on natural gas measurement.
However, because pulsation is a frequent cause
of gauge line error, it is important to review the
causes of pulsation induced SRE.
By: George L. Bell, Sr.
PGI International
Pulsation Testing Coordinator
1
Square Root Error
Most of the natural gas measurement in the United
States is performed with orifice plates.
The gas flow rate (Q) is calculated using the basic
formula Q = K√∆PXP. The fixed orifice coefficient (K)
is derived from a formula found in the latest edition of
AGA Report Number 3. Differential pressure (∆P)
and line pressure (P) are measured with mechanical
chart recorders or electronic transmitters.
Under steady-state flow conditions, gas volumes can
be accurately measured with current state-of-the-art
equipment. However, inaccurate measurement
occurs when the ∆P modulates, or changes, at a
frequency greater than the frequency that the
measurement system extracts the square root of the
∆P.
Pulsation from gas compressors, control valves,
pressure regulators, and some piping configurations
are one source of frequent ∆P modulation. Figure 1
illustrates the amplitude and frequency of pulsation
generated by a reciprocating compressor and a
control valve.
Figure 1

The compressor is causing the pulsation
between 0 and 50 Hz, and a control valve is
creating the pulse at 100 Hz.
The measurement error referred to above is
called Square Root Error (SRE). It is the
calculation of unsteady flow using the square
root of the average ∆P verses the average of
the square root values of the instantaneous ∆P.
Because SRE is directly related to flow
measurement error, it is a very important topic
to those who buy and sell natural gas.
In a paper presented at the 1989 Gulf Coast
Gas Measurement Short Course in-Houston,
Texas titled "Pulsation Effects on Orifice
Metering Considering Primary and Secondary
Elements", Mr. Robert J. McKee of SWRI
states, “a mathematical definition of SRE can
be developed from the fact that the average
flow is proportional to the Avg.√∆P as opposed
to the √Avg. ∆P!'”
The formula developed by SWRI to determine
the severity of SRE is illustrated below:
%SRE = (√Avg. ∆P - Avg.√∆P) x 100
Avg. √∆P
An illustration of a %SRE calculation of a series
of six ∆P readings taken from a pulsating flow is
shown below:
Reading ∆P’s (in.wc.) √∆P (in. wc.)
1 75.0 8.660
2 30.0 5.477
3 47.5 6.892
4 52.2 7.225
5 26.1 5.109
6 71.6 8.462
Total: 302.4 41.825
Avg.∆P: 50.4
√Avg.∆: 7.099 6.971
%SRE = (7.099 – 6.971) x 100 = +1.84%SRE
6.971
Note that %SRE is a positive number;
mathematically it is impossible for %SRE to be
2
negative. In addition, %SRE increases with pulsation
amplitude, and is inversely proportional to ∆P (for a
given pulsation amplitude %SRE will be greater at low
∆P than at high ∆P).
Other Primary Element Errors
Though SRE is the largest component of pulsation
induced primary element error, under extreme
pulsation conditions inertial error and coefficient shift
will both increase in magnitude. A brief explanation of
each follows:
Inertial Error: Pulsating gas flow will tend to remain in
motion due to its inertia. As a result, flow velocity
changes lag behind ∆P changes. Inertial errors are
insignificant unless pulsation amplitude and
frequency are both relatively high.
Coefficient Shifts: Though difficult to quantify, test
data indicates that pulsation levels above 1.5% SRE
contribute to shifts in the orifice co-efficient.
Measuring %SRE
When the %SRE at operating conditions is quantified,
it can be used to approximate the primary element
error induced by pulsation and to determine whether
or not corrective action is necessary.
Percent square root error is measured with a device
called the Square Root Error Indicator produced by
PGI International. This analytical instrument utilizes a
high frequency response ∆P transducer and special
software to calculate %SRE according to the formula
developed by SWRI shown earlier.
The PGI SRE Indicator is an analytical instrument
used to determine the severity of pulsation and
calculate %SRE. Because other primary element
errors (inertial error and co-efficient shifts) are not
measured, %SRE should not be used to correct flow
measurement readings, but to determine if corrective
action is necessary.
Reducing Pulsation
Because of the potential measurement error, many
buyers will not accept delivery of natural gas from
producers if the %SRE exceeds specific contract
limits. %SRE allowed by contract may vary

depending on daily volume, and could be as low
as 0.20 %SRE.
The simplest method of reducing pulsation
induced SRE is to raise the ∆P by changing the
orifice plate. Unfortunately, this may also limit
the operating range of the measurement
system.
In some cases, the piping system can be
modified or the pulsation source moved to
reduce SRE. This can be time consuming and
costly.
Another popular "cure' for high SRE is to install
a device, such as a restricting orifice, between
the pulsation source and the measuring station.
The negative effects of restricting devices are
increased cost of compression and a limited
flow range.
%SRE can also be reduced by installing an
acoustic filter to remove most of the pulsation.
Though more costly than a restricting device, a
properly designed acoustic filter will operate
over a much wider flow range with a lower
pressure drop.
If %SRE is above your contract limits, it is a
good idea to contact an expert such as
SouthWest Research in San Antonio, Texas for
more extensive testing to determine a solution.
Gauge Line Errors
Gauge Line Error (GLE) exists when the ∆P at
the taps does not equal the ∆P at the end of
the gauge lines. GLE may be caused by
pulsation or flow phenomena.
The gauge line starts at the orifice taps and
ends at the transmitter, flow computer, or
chart recorder connections. It includes any
pipe fittings, valves, valve manifolds, tube
fittings, instrument tubing, and condensate
chambers or bottles that may be installed
between the orifice taps and the measurement
device.
Based upon research conducted by SWRI, it
has been determined that gauge line error has
two components:
1) Gage Line Distortion: defined as the
amplification (increase) or attenuation
3
(decrease) of the pulsation amplitude in the
gauge lines. Gauge line distortion is similar to
the noise created when playing a flute or
blowing across the top of an open bottle.
2) Gauge Line Shift: defined as the actual shifting
of the average pressure along the length of the
gauge line. Gage line shift may be created by
pulsation rectification effects, which occur
when the ∆P signal is transmitted through
changing inside diameters of gauge line
created by multiple instrument tubing sizes,
bottles or condensate chambers, shut-off
valves with a smaller ID than the instrument
tubing, or male-to-female NPT connections.
Other causes of gauge line shift include gas
oscillation at the mouth of the gauge line, and
density changes caused by pressure and
temperature fluctuations in the gauge line.
Measuring Gauge Line Errors
In 1990, a GLE indicator was developed by PGI
International. In 1996, improvements to the earlier
model included the capability to perform both %SRE
and GLE tests. (shown in Figure 2). The Indicator
was developed by PGI International to indicate and
quantify both gauge line error and square root error.
Used extensively over the past five years, it has
proven to be quite useful in the study of gauge line
and square root error.
Figure 2
The GLE indicator is designed to compare the, ∆P at
the orifice taps to the ∆P at the end of the gauge
lines. If the two signals are not the same, the
difference will be due to gauge line error.

The GLE Indicator consists of two (2)
Rosemount 3051C "SMART" ∆P transmitters
and the necessary mounting hardware to install
one 3051 C directly at the orifice taps, and the
other at the end of the gauge lines.
A PCMCIA A/D converter changes the analog
output signals from each 3051C transmitter into
a digital signal. Which is then analyzed by
special software in a portable, battery operated
lap top computer. All software is Windows and
Windows 95 based for ease of operation.
Operation of the GLE Indicator is
straight forward. Station and test data are
entered as shown in Figures 3 and 4.
Figure 3
Figure 4
4
A calibration procedure corrects any output signal
deviations at shown input differential pressures (see
Figure 5).
Figure 5
A "noise test" is then performed to determine system
error (see Figure 6). During the "noise test,' both
3051 C's measure the same differential pressure,
which is manually modulated using a hand pump.
The noise test will show that the 3051C's and the
Validyne are reading the same and no outside
electrical interference is present.
Figure 6

The 3051 C transmitters are then installed as in
Figure 2, pressurized to line pressure and rezeroed
to correct for transmitter position and
static pressure zero shift.
During the timed GLE test, the ∆P of each
3051C is sampled at a frequency of 250
times/second. A flow calculation for each
transmitter is performed once each second
using the average of the ∆P's.
The GLE Indicator graphically indicates the true
∆P at the orifice and the actual gauge line error.
Average gauge line error in inches of water and
%GLE are also displayed. At the conclusion of
the test, the flow calculated from each
transmitter is compared. Any difference in thetwo
volumes is displayed in annualized MMCF
of gas and dollars of gain or loss. The dollar
amount is calculated based on the user’s price
of gas per MCF at the time of the test.
Numerous GLE tests can be performed at
various flow rates. All gauge line and square
root error tests are automatically stored to the
hard disk for future analysis or printing.
Testing Results
Extensive field testing with the GLE Indicator
confirmed the research conducted at SWRI by
PCRC. In the following test examples, "DM"
represents the ∆P of the reference 3051C that
was direct mounted on one side of the orifice
fitting.
GLE-Test 1: In this test, gauge line error was
measured at the outlet of a small bore (0.187")
5 valve manifold connected to the orifice fitting
with 4.5 ft. of 3/8" O.D. tubing. A pair of 3/8’ I.D.
ball valves was installed at the taps. %SRE at
the taps was .997% as shown in Figure 1.
Gauge line error was negative 0.102 wc. at
30.7 inches wc. ∆P. Annualized, this would
result in a measurement error of $31,503.00 as
shown in Figure 7.
5
Figure 7
GLE-Test 2: In this test, gauge line error was
measured at the end of the 3/8" O.D. gauge lines 5 ft.
from the orifice taps. Block valves on the 6" Daniel
Senior Fitting were conventional 0. 1 87" soft seat
needle valves. GLE measured negative 0.386 in. wc.
at a constant 36.07 MMCFD flow rate. Based on gas
at $2.00 per MCF, the annualized error was a
negative $132,309.00 as shown in Figure 8.
Figure 8

GLE-Test 3: A Honeywell ∆P transmitter was
connected to a 16' Daniel Junior fitting with 15
ft. of 3/8" O.D. tubing, a small bore (0.187") 5
valve manifold, and 1/2" ball valves at the orifice
taps. At a constant flow of 30.19 MMCFD,
gauge line error measured a negative 0.192 in.
wc. Annualized this would result in a
measurement error of $143,529.00 as shown
below in Figure 9.
Figure 9
Eliminating or Minimizing GLE
As mentioned previously, numerous gas
contracts now include pulsation magnitude
clauses. Many transmission companies
require the installation of acoustic filters to
minimize pulsation levels and %SRE.
Testing has shown that in some cases
gauge line error may continue to be present
after the installation of an acoustic filter and
when % of SRE readings are very low - in
some cases 0.1 % SRE.
Because of the many variables involved in
gauge line error, such as pulsation levels,
gauge line lengths and/or diameters,
operating pressure, gas density, and gas
velocity, it is extremely difficult to observe a
measurement location and predict what
gauge line error, if any, will be present. The
only way to determine the presence of gauge
line error is through testing.
An alternative to testing for gauge line error
at each installation is to install the transmitter
6
and/or EFM in a manner that will minimize or
eliminate gauge line error.
To minimize gauge line error when pulsation or flow
phenomena are present, a differential measuring
device should be close coupled to orifice fittings with
equal lengths of large bore (0.375" I.D. or greater)
constant diameter gauge lines.
This can be accomplished using a short length of 1/2"
O.D. instrument tubing and full opening ball valves on
the orifice fitting and measurement device. However,
in this type of installation the numerous matting of
female NPT connections create small 'volume
chambers," which could produce gauge line shift
(pulsation rectification effects).
If the full port 1" ball valves are used, it is very easy to
shock one side of the measurement device with full
line pressure. When this occurs, it may create a
significant static shift in the calibration of the
transmitter not detectable under normal calibration
procedures.
A more feasible method is the Direct Mounting
System as shown in Figure 10. With over 10,000
installations currently in service, this method of close
coupling transmitters or a self-contained EFM to the
orifice fitting continues to gain wide acceptance within
the industry.
Utilizing a pair of stabilized connectors for safety, and
roddable large bore (0.375" I.D.) soft seat instrument
manifolds, the transmitters are located as close as

possible to the orifice taps, typically within 13
inches or less.
Flanged connections with Teflon® seal’s
eliminate the "volume chambers" found with
NPT connections. Multiple turn valves prevent
"shocking" of the measurement device when
opened.
Frequent valve stem packing adjustments are
eliminated by a dynamically loaded stem seal
that is guaranteed leak-free under fluctuating
pressure (vacuum to +10,000 PSI) and
temperature (-40° to +450°F).
Protection of the electronic transmitters from
cathodic protection currents and possible
transients is provided through the use of
dielectric isolators, rated to 2,500 volts DC,
located between the stabilized connectors and
the manifold.
Summary
• Pulsation created by compressors, flow
control valves, regulators, and some piping
configurations may create unacceptable
levels of SRE and/or GLE.
• %SRE is always positive, never negative,
and increases with pulsation amplitude.
• To reduce pulsation to an acceptable level,
the source of the pulsation must be
eliminated, piping systems modified, ∆P
increased, a restricting device installed, or a
properly sized acoustic filter installed.
• Whenever pulsation is present there is a
high probability that GLE will exist.
• Even though %SRE readings may be within
acceptable limits, low amplitude/high
frequency pulsations may create significant
GLE.
• Volume chambers, such as condensate
pots, bottles, and bellows housings of
different volumes greatly affect GLE.
• Numerous measurement devices
connected to the same set of orifice taps
may create GLE.
7
• Unlike %SRE, GLE can be either positive or
negative. It may be of a greater magnitude than
%SRE.
• To minimize or eliminate GLE, the transmitters or
EFM should be close coupled to the orifice taps
with equal length, large bore (0.375" I.D. or
greater), constant diameter gauge lines.
• Minimizing or eliminating GLE by close coupling
the measurement device to the orifice taps will
not reduce %SRE.
References
1. McKee, Robert S. 'Pulsation Effects on Orifice
Metering Considering Primary and Secondary
Elements," Proceedings of the Twenty-Second
Gulf Coast Measurement Short Course, pp.
112-118, 1989.
2. Gegg, Debbie, "Effects and Control of
Pulsation's in Gas Measurement," Arkia
Energy Resources, Shreveport, La.
Proceedings of the Sixty-Fourth International
School of Hydrocarbon Measurement, pp. 331-
335, 1989.
3. Durke, Ray G. and Sparks, Cecil R., "Pulsation
and Transient-Induced Errors at Orifice Meter
Installations,' PCRC Technical Report No. 87-
3, 1987.
4. Floyd, J.H. and Everett, W.S. "The Effect of
High Frequency Pulsation's on Differential
Meter
Accuracy," Proceedings of the Twenty Third
Appalachian Gas Measurement Short Course,
pp. 283-294, 1963.
5. Durke Ray G., Smalley, Anthony J., and
McKee, Robert J. "An Assessment of
Technology for Correcting Pulsation-Induced
Orifice Flow Measurement Errors," PCRC
Technical Report No. TA 91-1, 1991.